GM Joel Benjamin vs Leela Knight Odds Classical match

[Uri Blass]

Good afternoon Mr. Larry Kaufman.
For me, it continues to be very enjoyable and an object of teaching and learning, watching LeelaQueenOdds' games against different levels of players. To tell the truth, against all types of human players, I am surprised by the capacity of the computer. I don't understand what has happened to the table of the 100 best humans, it has not been updated for a long time, it seems as if interest in it has been lost. On the computer at my desk in tournament mode I will soon be playing against different engines in full conditions and at the minute. I hope to update my knowledge a little to the current state of the art. Happy afternoon I wish you.

Marcus, who maintinas the list of top 100 at queen odds, has been busy with university exams, but perhaps that is changing and he might have time to resume the project. I can tell you that if the list were updated now, GM Joel Benjamin, handle "SuperEasy", would show a large jump up, far above everyone else, since he continues to play the bot, now at a reduced 8'3" instead of 10'3", and is winning about 95% of the games that way! Clearly he still plays at GM level, as I don't think anyone below GM level (regardless of actual title) would be able to score that well at that time control. One thing it demonstrates is the enormous difference in level of play between Rapid and bullet chess. Of course we all knew that there was a big difference, but how big was hard to say. Top bullet specialists often give 5 minutes to 1 minute odds to players around IM level or so, but that is misleading because they think on the opponent's time. What the bot games show is that a GM level blitz player (Joel is 2520 FIDE blitz after the recent World Blitz Championship) can beat the bot by about 500 elo at fast Rapid, whereas at 1'0" bullet even much stronger blitz players (including Wei Yi, 2711 FIDE blitz, lose by more than 200 elo (Leela always White), so the gap in quality between fast Rapid and 1'0" bullet appears to be about 900 elo points! Meanwhile the knight odds bot with current net now has a total blitz (3'2" minimum) record against the world's top fifty FIDE classical players of 64 wins, 26 draws, and ZERO losses!! We are continuing to work on the bots, there is still ample room for improvement, especially for the queen odds bot even though it is already amazingly strong.

I consider the conslusion to be wrong because queen odds is not normal game and people may be relatively significantly stronger in queen odds if they train a lot in this way.

I do not know how many FM or IM level players even tried 8+3 or slower time control against the queen odds bot(usually they play faster time control).

[Father]

Good afternoon Mr. Larry Kaufman.
For me, it continues to be very enjoyable and an object of teaching and learning, watching LeelaQueenOdds' games against different levels of players. To tell the truth, against all types of human players, I am surprised by the capacity of the computer. I don't understand what has happened to the table of the 100 best humans, it has not been updated for a long time, it seems as if interest in it has been lost. On the computer at my desk in tournament mode I will soon be playing against different engines in full conditions and at the minute. I hope to update my knowledge a little to the current state of the art. Happy afternoon I wish you.

Marcus, who maintinas the list of top 100 at queen odds, has been busy with university exams, but perhaps that is changing and he might have time to resume the project. I can tell you that if the list were updated now, GM Joel Benjamin, handle "SuperEasy", would show a large jump up, far above everyone else, since he continues to play the bot, now at a reduced 8'3" instead of 10'3", and is winning about 95% of the games that way! Clearly he still plays at GM level, as I don't think anyone below GM level (regardless of actual title) would be able to score that well at that time control. One thing it demonstrates is the enormous difference in level of play between Rapid and bullet chess. Of course we all knew that there was a big difference, but how big was hard to say. Top bullet specialists often give 5 minutes to 1 minute odds to players around IM level or so, but that is misleading because they think on the opponent's time. What the bot games show is that a GM level blitz player (Joel is 2520 FIDE blitz after the recent World Blitz Championship) can beat the bot by about 500 elo at fast Rapid, whereas at 1'0" bullet even much stronger blitz players (including Wei Yi, 2711 FIDE blitz, lose by more than 200 elo (Leela always White), so the gap in quality between fast Rapid and 1'0" bullet appears to be about 900 elo points! Meanwhile the knight odds bot with current net now has a total blitz (3'2" minimum) record against the world's top fifty FIDE classical players of 64 wins, 26 draws, and ZERO losses!! We are continuing to work on the bots, there is still ample room for improvement, especially for the queen odds bot even though it is already amazingly strong.

Thank you Mr. Larry Kaufman. For me, the fact of being where I am is already a reward of life. Furthermore, I will undoubtedly arrive at the time, place and position that our Programmer wants.
Wonderful that GM Joel Benjamin leads the table, I have not seen Uri play with LeelaQueenOdds again and yesterday I observed that you have not played against LeelaQueenOdds for several weeks.
The standings give more excitement to everything, I observe a large number of players in combat against the computer, the street ring is still completely valid: Great idea!!

[lkaufman]

Good afternoon Mr. Larry Kaufman.
For me, it continues to be very enjoyable and an object of teaching and learning, watching LeelaQueenOdds' games against different levels of players. To tell the truth, against all types of human players, I am surprised by the capacity of the computer. I don't understand what has happened to the table of the 100 best humans, it has not been updated for a long time, it seems as if interest in it has been lost. On the computer at my desk in tournament mode I will soon be playing against different engines in full conditions and at the minute. I hope to update my knowledge a little to the current state of the art. Happy afternoon I wish you.

Marcus, who maintinas the list of top 100 at queen odds, has been busy with university exams, but perhaps that is changing and he might have time to resume the project. I can tell you that if the list were updated now, GM Joel Benjamin, handle "SuperEasy", would show a large jump up, far above everyone else, since he continues to play the bot, now at a reduced 8'3" instead of 10'3", and is winning about 95% of the games that way! Clearly he still plays at GM level, as I don't think anyone below GM level (regardless of actual title) would be able to score that well at that time control. One thing it demonstrates is the enormous difference in level of play between Rapid and bullet chess. Of course we all knew that there was a big difference, but how big was hard to say. Top bullet specialists often give 5 minutes to 1 minute odds to players around IM level or so, but that is misleading because they think on the opponent's time. What the bot games show is that a GM level blitz player (Joel is 2520 FIDE blitz after the recent World Blitz Championship) can beat the bot by about 500 elo at fast Rapid, whereas at 1'0" bullet even much stronger blitz players (including Wei Yi, 2711 FIDE blitz, lose by more than 200 elo (Leela always White), so the gap in quality between fast Rapid and 1'0" bullet appears to be about 900 elo points! Meanwhile the knight odds bot with current net now has a total blitz (3'2" minimum) record against the world's top fifty FIDE classical players of 64 wins, 26 draws, and ZERO losses!! We are continuing to work on the bots, there is still ample room for improvement, especially for the queen odds bot even though it is already amazingly strong.

Thank you Mr. Larry Kaufman. For me, the fact of being where I am is already a reward of life. Furthermore, I will undoubtedly arrive at the time, place and position that our Programmer wants.
Wonderful that GM Joel Benjamin leads the table, I have not seen Uri play with LeelaQueenOdds again and yesterday I observed that you have not played against LeelaQueenOdds for several weeks.
The standings give more excitement to everything, I observe a large number of players in combat against the computer, the street ring is still completely valid: Great idea!!

When I have time, I do play the bots, usually in Rapid, but I like to play at odds that are difficult but not impossible for me. Fair odds in fast Rapid for me seem to be two knights (using the "LeelaPieceOdds") bot, so i play that sometimes. Or I play at Rook odds or Rook and move odds, where I am often winning for a long time but usually find a way to lose. At least I feel that I should be able to win some games with practice and care, at least at Rook and move odds.

[Father]

Good afternoon Mr. Larry Kaufman.
For me, it continues to be very enjoyable and an object of teaching and learning, watching LeelaQueenOdds' games against different levels of players. To tell the truth, against all types of human players, I am surprised by the capacity of the computer. I don't understand what has happened to the table of the 100 best humans, it has not been updated for a long time, it seems as if interest in it has been lost. On the computer at my desk in tournament mode I will soon be playing against different engines in full conditions and at the minute. I hope to update my knowledge a little to the current state of the art. Happy afternoon I wish you.

Marcus, who maintinas the list of top 100 at queen odds, has been busy with university exams, but perhaps that is changing and he might have time to resume the project. I can tell you that if the list were updated now, GM Joel Benjamin, handle "SuperEasy", would show a large jump up, far above everyone else, since he continues to play the bot, now at a reduced 8'3" instead of 10'3", and is winning about 95% of the games that way! Clearly he still plays at GM level, as I don't think anyone below GM level (regardless of actual title) would be able to score that well at that time control. One thing it demonstrates is the enormous difference in level of play between Rapid and bullet chess. Of course we all knew that there was a big difference, but how big was hard to say. Top bullet specialists often give 5 minutes to 1 minute odds to players around IM level or so, but that is misleading because they think on the opponent's time. What the bot games show is that a GM level blitz player (Joel is 2520 FIDE blitz after the recent World Blitz Championship) can beat the bot by about 500 elo at fast Rapid, whereas at 1'0" bullet even much stronger blitz players (including Wei Yi, 2711 FIDE blitz, lose by more than 200 elo (Leela always White), so the gap in quality between fast Rapid and 1'0" bullet appears to be about 900 elo points! Meanwhile the knight odds bot with current net now has a total blitz (3'2" minimum) record against the world's top fifty FIDE classical players of 64 wins, 26 draws, and ZERO losses!! We are continuing to work on the bots, there is still ample room for improvement, especially for the queen odds bot even though it is already amazingly strong.

Thank you Mr. Larry Kaufman. For me, the fact of being where I am is already a reward of life. Furthermore, I will undoubtedly arrive at the time, place and position that our Programmer wants.
Wonderful that GM Joel Benjamin leads the table, I have not seen Uri play with LeelaQueenOdds again and yesterday I observed that you have not played against LeelaQueenOdds for several weeks.
The standings give more excitement to everything, I observe a large number of players in combat against the computer, the street ring is still completely valid: Great idea!!

When I have time, I do play the bots, usually in Rapid, but I like to play at odds that are difficult but not impossible for me. Fair odds in fast Rapid for me seem to be two knights (using the "LeelaPieceOdds") bot, so i play that sometimes. Or I play at Rook odds or Rook and move odds, where I am often winning for a long time but usually find a way to lose. At least I feel that I should be able to win some games with practice and care, at least at Rook and move odds.

February 10, 2025 "The small footprint of my wooden horse:" "Catecan."

Today is a very special day for me. And special days call us to give thanks. Thus:
"First of all" I thank the Programmer of my life, existence and human algorithm, the "Programmer of life." "Thank you for life".
"Secondly" I thank my deceased Dad and Mom, my family and my friends: "Thank you Dad and Mom for being "co-creators" of my life and opportunity to be and live. Thank you for teaching me how to play chess.
"Thirdly, thanks to the traditions and values ​​of the elders and ancestors" to my Father and his little wooden horse called "Catecan" and photographed in 1915, which has been my inspiration to fight to conquer the world of dreams and ideals in chess and computers, even if it is only inside a story book, if it is, "my purpose in life."
"Fourthly" I thank the Leela team and Mr. Larry Kaufman, Grandmaster and former Senior World Champion, for allowing me to play against the Odds robots and compete to be part of the ecosystem of the 100 best humans who face LeelaQueenOdds.
Fifthly. Thanks to the chess Masters who have allowed us players who compete in "Wooden Horses" to have been admitted to an open competition, it is very clear to me that the Masters and Professionals of chess are the salt and the main condiment that give the best and main flavor to the food, without them, the greatness of chess would be spoiled.
Me and Catecan occupying second place in the table, even if only for the moment, far exceeds any dream and ideal, beyond reality and higher than everything that fits and could be covered within a storybook.
Sicerely,
Pablo Ignacio Restrepo
Father; Grand-Father and Catecan in playchess
Catecan in Lichess
Father in talkchess

Leaderboard - Top 100
About this Leaderboard
How do I earn points? You can gain or lose points only by playing LeelaQueenOdds
How is the rating computed? It's based on the same method as the FIDE rating adjustment, using:
K = 40 for the first 30 games
K = 20 up to 150 games
K = 10 for subsequent games
Play at your best Time Control: This leaderboard adjusts ratings based on Time Control
How is Leela's rating computed? A statistical model based on LeelaQueenOdds games normalizes Lichess blitz ratings and time control with the score expected by Elo
Play for the Win: Draws have the 'K' factor halved
Starting rating: Lichess rating - 200, with a maximum starting rating of 1800
Update: In these early stages, the leaderboard is updated occasionally but daily
Github: https://github.com/Marcogio9/LeelaQueenOdds-Leaderboard
No Player Rating Games
LeelaQueenOdds as white 1+0 (Reference) 2895.2
LeelaQueenOdds as black 1+0 (Reference) 2795.2
1 SuperEasy 2657.3 230
LeelaQueenOdds as white 1+1 (Reference) 2634.5
LeelaQueenOdds as white 3+0 (Reference) 2619.5
LeelaQueenOdds as black 1+1 (Reference) 2534.5
LeelaQueenOdds as black 3+0 (Reference) 2519.5
2 Catecan 2475.6 4518
3 rowrulz 2454.6 82
4 Orichess88 2450.4 71
LeelaQueenOdds as white 3+2 (Reference) 2450.0
5 Hissha 2427.1 182
6 Master_Unknown 2414.6 92
7 crtex 2405.6 342
8 RhinoTank 2389.3 166
9 Chrisinc 2378.5 256
10 loeschick 2356.2 60
LeelaQueenOdds as white 5+3 (Reference) 2353.7
LeelaQueenOdds as black 3+2 (Reference) 2350.0
11 Omnivor 2328.4 40
12 Bonitta 2326.6 67
13 Ushkothebear 2292.7 632
14 AxelBoldt 2284.5 1013
15 edgebottle 2274.6 1498
16 Sakali 2259.6 230
LeelaQueenOdds as black 5+3 (Reference) 2253.7
17 cmdbreivik 2249.6 2510
18 CoolerMove 2243.1 23
19 Derrick3110 2233.7 107
20 Letsgo_Torch 2221.2 52
LeelaQueenOdds as white 10+5 (Reference) 2220.3
21 Respectful_Dave 2217.0 75
22 ricky8632200 2215.1 178
23 Zmof 2204.5 118
24 itore 2197.1 43
25 Kelvandil 2175.1 47
26 Peristilo 2168.7 1267
27 ysbeer 2165.4 313
28 jzm 2165.3 52
29 MindBoggle 2165.2 43
30 Feofanov_Vladimir 2164.7 143
31 DenishevM 2161.9 31
32 uriblass 2154.3 227
33 HawkingFeynman 2147.2 17
34 Ajktulm 2138.0 657
35 yyzx77 2136.5 32
36 lx_chessmath 2129.9 63
37 Spreek 2129.8 45
LeelaQueenOdds as white 15+10 (Reference) 2127.3
38 SilverMitt 2123.1 26
39 DubiousMistakes 2120.8 47
40 chess66666 2120.5 524
LeelaQueenOdds as black 10+5 (Reference) 2120.3
41 brewmasterua 2114.3 45
42 nn1133 2112.1 6418
43 sfy2011 2111.8 37
44 avemaria2025 2109.6 109
45 pzpzzz 2107.1 80
46 WhiteRabbit00735 2105.5 210
47 ezitis_migla 2104.8 135
48 iWannaSeeYouSmile 2099.8 73
49 Puzzleman40 2096.3 33
50 kbll 2084.4 102
51 wyw19901990 2077.0 45
52 scrooge 2065.4 100
53 wael96 2061.0 53
54 Olga547823 2053.3 33
55 RabbitsGrowOnTrees 2049.9 297
56 Sir_Patzer 2046.1 14
57 Forgive1219 2044.7 189
58 TheLlamaLord 2043.7 22
59 elputoano 2041.4 188
60 lyxsgdzt 2040.1 66
61 qasdfgi 2038.4 63
62 xHadyx 2035.3 182
63 mapascual 2030.3 42
64 AndrewHardegen 2027.7 67
65 chutop 2027.4 9
LeelaQueenOdds as black 15+10 (Reference) 2027.3
66 Jagen 2024.5 56
67 Fred_Fertig_64 2023.8 24
68 Loks71 2012.3 39
69 Monchi 2009.0 29
70 winthlin 2005.3 27
71 OTBquadsicles 2003.4 10
72 o4epegHou 1998.4 84
73 A-Liang 1994.3 36
74 jimmydeeznuts 1991.6 820
75 universe64 1988.7 39
76 paCOOLa 1987.1 20
77 Kostik_Mostik 1987.0 20
78 MathTennisCoding 1982.7 99
79 AdmiralHabibi 1978.5 77
80 cramor 1977.8 42
81 invab 1977.2 42
82 gambitnobueno 1974.7 20
83 Lukas-2010 1972.3 30
84 dzxyyds 1969.3 206
85 onemillonprime 1969.2 161
86 machineplayer9 1968.5 22
87 T_Sapphirine_Kisaki 1967.2 46
88 LeopardCheetah 1959.1 18
89 Abigail_Morgan 1959.0 10
90 gatsten 1958.8 11
91 Andromeda3 1957.6 203
92 lolotron 1957.3 34
93 adirac 1956.2 49
94 Punypawn315 1954.9 19
95 wireless78 1954.4 21
96 Pulevik2022 1953.4 96
97 kuromiko 1951.2 102
98 frums321 1948.7 15
99 mla96 1947.5 106
100 Liouville1 1940.3 42

[lkaufman]

February 10, 2025 "The small footprint of my wooden horse:" "Catecan."

Today is a very special day for me. And special days call us to give thanks. Thus:
"First of all" I thank the Programmer of my life, existence and human algorithm, the "Programmer of life." "Thank you for life".
"Secondly" I thank my deceased Dad and Mom, my family and my friends: "Thank you Dad and Mom for being "co-creators" of my life and opportunity to be and live. Thank you for teaching me how to play chess.
"Thirdly, thanks to the traditions and values ​​of the elders and ancestors" to my Father and his little wooden horse called "Catecan" and photographed in 1915, which has been my inspiration to fight to conquer the world of dreams and ideals in chess and computers, even if it is only inside a story book, if it is, "my purpose in life."
"Fourthly" I thank the Leela team and Mr. Larry Kaufman, Grandmaster and former Senior World Champion, for allowing me to play against the Odds robots and compete to be part of the ecosystem of the 100 best humans who face LeelaQueenOdds.
Fifthly. Thanks to the chess Masters who have allowed us players who compete in "Wooden Horses" to have been admitted to an open competition, it is very clear to me that the Masters and Professionals of chess are the salt and the main condiment that give the best and main flavor to the food, without them, the greatness of chess would be spoiled.
Me and Catecan occupying second place in the table, even if only for the moment, far exceeds any dream and ideal, beyond reality and higher than everything that fits and could be covered within a storybook.
Sicerely,
Pablo Ignacio Restrepo
Father; Grand-Father and Catecan in playchess
Catecan in Lichess
Father in talkchess

Leaderboard - Top 100
About this Leaderboard
How do I earn points? You can gain or lose points only by playing LeelaQueenOdds
How is the rating computed? It's based on the same method as the FIDE rating adjustment, using:
K = 40 for the first 30 games
K = 20 up to 150 games
K = 10 for subsequent games
Play at your best Time Control: This leaderboard adjusts ratings based on Time Control
How is Leela's rating computed? A statistical model based on LeelaQueenOdds games normalizes Lichess blitz ratings and time control with the score expected by Elo
Play for the Win: Draws have the 'K' factor halved
Starting rating: Lichess rating - 200, with a maximum starting rating of 1800
Update: In these early stages, the leaderboard is updated occasionally but daily
Github: https://github.com/Marcogio9/LeelaQueenOdds-Leaderboard
No Player Rating Games
LeelaQueenOdds as white 1+0 (Reference) 2895.2
LeelaQueenOdds as black 1+0 (Reference) 2795.2
1 SuperEasy 2657.3 230
LeelaQueenOdds as white 1+1 (Reference) 2634.5
LeelaQueenOdds as white 3+0 (Reference) 2619.5
LeelaQueenOdds as black 1+1 (Reference) 2534.5
LeelaQueenOdds as black 3+0 (Reference) 2519.5
2 Catecan 2475.6 4518
3 rowrulz 2454.6 82
4 Orichess88 2450.4 71
LeelaQueenOdds as white 3+2 (Reference) 2450.0
5 Hissha 2427.1 182
6 Master_Unknown 2414.6 92
7 crtex 2405.6 342
8 RhinoTank 2389.3 166
9 Chrisinc 2378.5 256
10 loeschick 2356.2 60
LeelaQueenOdds as white 5+3 (Reference) 2353.7
LeelaQueenOdds as black 3+2 (Reference) 2350.0
11 Omnivor 2328.4 40
12 Bonitta 2326.6 67
13 Ushkothebear 2292.7 632
14 AxelBoldt 2284.5 1013
15 edgebottle 2274.6 1498
16 Sakali 2259.6 230
LeelaQueenOdds as black 5+3 (Reference) 2253.7
17 cmdbreivik 2249.6 2510
18 CoolerMove 2243.1 23
19 Derrick3110 2233.7 107
20 Letsgo_Torch 2221.2 52
LeelaQueenOdds as white 10+5 (Reference) 2220.3
21 Respectful_Dave 2217.0 75
22 ricky8632200 2215.1 178
23 Zmof 2204.5 118
24 itore 2197.1 43
25 Kelvandil 2175.1 47
26 Peristilo 2168.7 1267
27 ysbeer 2165.4 313
28 jzm 2165.3 52
29 MindBoggle 2165.2 43
30 Feofanov_Vladimir 2164.7 143
31 DenishevM 2161.9 31
32 uriblass 2154.3 227
33 HawkingFeynman 2147.2 17
34 Ajktulm 2138.0 657
35 yyzx77 2136.5 32
36 lx_chessmath 2129.9 63
37 Spreek 2129.8 45
LeelaQueenOdds as white 15+10 (Reference) 2127.3
38 SilverMitt 2123.1 26
39 DubiousMistakes 2120.8 47
40 chess66666 2120.5 524
LeelaQueenOdds as black 10+5 (Reference) 2120.3
41 brewmasterua 2114.3 45
42 nn1133 2112.1 6418
43 sfy2011 2111.8 37
44 avemaria2025 2109.6 109
45 pzpzzz 2107.1 80
46 WhiteRabbit00735 2105.5 210
47 ezitis_migla 2104.8 135
48 iWannaSeeYouSmile 2099.8 73
49 Puzzleman40 2096.3 33
50 kbll 2084.4 102
51 wyw19901990 2077.0 45
52 scrooge 2065.4 100
53 wael96 2061.0 53
54 Olga547823 2053.3 33
55 RabbitsGrowOnTrees 2049.9 297
56 Sir_Patzer 2046.1 14
57 Forgive1219 2044.7 189
58 TheLlamaLord 2043.7 22
59 elputoano 2041.4 188
60 lyxsgdzt 2040.1 66
61 qasdfgi 2038.4 63
62 xHadyx 2035.3 182
63 mapascual 2030.3 42
64 AndrewHardegen 2027.7 67
65 chutop 2027.4 9
LeelaQueenOdds as black 15+10 (Reference) 2027.3
66 Jagen 2024.5 56
67 Fred_Fertig_64 2023.8 24
68 Loks71 2012.3 39
69 Monchi 2009.0 29
70 winthlin 2005.3 27
71 OTBquadsicles 2003.4 10
72 o4epegHou 1998.4 84
73 A-Liang 1994.3 36
74 jimmydeeznuts 1991.6 820
75 universe64 1988.7 39
76 paCOOLa 1987.1 20
77 Kostik_Mostik 1987.0 20
78 MathTennisCoding 1982.7 99
79 AdmiralHabibi 1978.5 77
80 cramor 1977.8 42
81 invab 1977.2 42
82 gambitnobueno 1974.7 20
83 Lukas-2010 1972.3 30
84 dzxyyds 1969.3 206
85 onemillonprime 1969.2 161
86 machineplayer9 1968.5 22
87 T_Sapphirine_Kisaki 1967.2 46
88 LeopardCheetah 1959.1 18
89 Abigail_Morgan 1959.0 10
90 gatsten 1958.8 11
91 Andromeda3 1957.6 203
92 lolotron 1957.3 34
93 adirac 1956.2 49
94 Punypawn315 1954.9 19
95 wireless78 1954.4 21
96 Pulevik2022 1953.4 96
97 kuromiko 1951.2 102
98 frums321 1948.7 15
99 mla96 1947.5 106
100 Liouville1 1940.3 42

Congratulations on second place! All of the top dozen except for you and number 10, are either known to be GM, IM, or FM, or have ratings and/or puzzle results that strongly suggest that they have these titles or at least the strength typical of those titles. Your approach is unique, aiming for draws in bullet rather than wins in blitz or Rapid, and it clearly works for you. GM Joel Benjamin ("SuperEasy") deserves special acknowledgement for being over 180 elo above all others, including many titled players. Moreover he did this without benefitting from the weakness of the bot when playing Black, due to having been trained only to play White. I did a survey of the performance ratings in rapid and in blitz of the bot since getting its own net, and the difference is about 150 elo, whereas it is only counted as 100 elo. I think this accomplishment makes the narrow 3.5 to 4.5 loss for LeelaKnightOdds to Joel at a Classical Time limit look pretty good, clearly he understands how to exploit handicaps quite well. I expect that his rating on this list will continue to rise if he keeps playing, since his results have been improving and the rating hasn't caught up to his recent results yet.

[Uri Blass]

results of Joel Benjamin against the queen odd bot:

2.12.2024-3.12.2024 when I search in lichess means 2.12.2024

2.12.2024 2-1(loss in the second game)
3.12.2024 3-0
4.12.2024 4.5-0.5(draw in the last game)
9.12.2024 7-0
10.12.2024 2.5-0.5(draw in last game)
11.12.2024 2-0
13.12.2024 2-0
14.12.2024 2-0
16.12.2024 4.5-0.5(draw in the second game)
17.12.2024 6-0
23.12.2024 2-1(loss first game)
31.12.2025 1-0
1.1.2025 1-0
3.1.2025 6-0
4.1.2025 3-0
5.1.2025 12-0
6.1.2025 7-0
7.1.2025 9.5-0.5(draw in game 9) had a record of 40 wins in a row from 24.12.2024 to 8.1.2025
11.1.2025 4-0
12.1.2025 3-1(lost game 3)
13.1.2025 4-1(lost game 2)
14.1.2025 6-1(lost game 6)
15.1.2025 1-0
16.1.2025 2-0
17.1.2025 2-0
18.1.2025 1-0
20.1.2025 4-0
21.1.2025 5-0
22.1.2025 1-0
23.1.2025 7-0
25.1.2025 2.5-0.5(draw in the second game)
28.1.2025 2-0
29.1.2025 12-0
30.1.2025 7-0
31.1.2025 6-1(lost game 6 on time)
1.2.2025 3-0
2.2.2025 9-0 started 8+3 in that day
3.2.2025 10-1 lost game 3
4.2.2025 2-0
5.2.2025 8-1 lost game 4
6.2.2025 14-0
7.2.2025 4-0
8.2.2025 4-1 lost first game
9.2.2025 12.5-0.5 drew game 9
10.12.2025 9-0
11.12.2025 10-0
12.12.2025 5-0
13.12.2025 2-0
so far 30 games in a row at 8+3 when the record at 10+3 was 40 games in a row

[jp]

...
Final game was a miracle draw by Leela in a losing endgame...

Can someone please post the moves of this miracle draw here?

... 82... Re8?? { [%eval -0.16] } { Blunder. Rd6 was best. } { [%clk 0:04:42] } (82... Rd6 83. Rg2 Nb3 84. Ke5 Rc6 85. Rb2 Kb4 86. Rh2 Rg6 87. Kf5 Ra6 88. Kg4) 83. Kg6 { [%eval -0.14] [%clk 1:35:27] } ...

Thanks for posting the pgn, Uri.